Original Article
Is 24 hour observation in hospital
after stopping intravenous antibiotics
in neonates justified?
Josella Fenech, Hermoine Andrejevic, Valerie Said Conti, Simon Attard Montalto
Abstract
Background: Antibiotics are given empirically for suspected
sepsis in up to 75% of neonates on the Neonatal and Paediatric
Intensive Care Unit (NPICU), after completion of a septic screen.
Treatment is discontinued on day 3 if cultures remain negative
or after 7-14 days with proven sepsis and, until recently, these
neonates are then observed for an additional period of 24 hours
before being discharged from hospital.
Aim: To assess whether the 24 hour observation period after
stopping antibiotics is clinically justified and, if not, whether
neonates can be discharged safely on the same day when
antibiotics are stopped.
Methods: A consecutive sample of 95 babies admitted to
NPICU, and who received antibiotics, from December 2006
to January 2008 were analysed prospectively. Their clinical
presentation, predisposing risk factors for neonatal sepsis,
investigations, antibiotic details and medical management
including respiratory support were recorded, and correlated
with all events that may have occurred during the observation
period after stopping antibiotics.
Keywords
Antibiotics, neonates, hospital observation
Josella Fenech MD, MRCPCH
Department of Paediatrics, Mater Dei Hospital, Tal-Qroqq, Malta
Hermoine Andrejevic MD MRCPCH
Department of Paediatrics, Mater Dei Hospital, Tal-Qroqq, Malta
Valerie Said Conti MD, MRCPCH
Department of Paediatrics, Mater Dei Hospital, Tal-Qroqq, Malta
Simon Attard Montalto* MD (L’pool), FRCPCH
Department of Paediatrics, Mater Dei Hospital, Tal-Qroqq, Malta
Email: simon.attard-montalto@gov.mt
* corresponding author
24
Results: No adverse events were documented in the 24
hour period after antibiotics in all 95 neonates in this study
and, therefore, there was no association with any potential
predisposing risk factors.
Conclusion: The need to observe neonates for a period
prior to discharge after stopping antibiotics is not supported
on clinical grounds and, as a result of this study, has been
discontinued. Neonates can be discharged from hospital safely
and immediately on stopping antibiotics, thus reducing hospital
stay and an estimated cost saving of approximately €18,000 to
the service provider per annum.
Introduction
Neonatal sepsis is classified into early onset (infection
occurring in the first 5 days of life) and late onset (after 5 days of
age).1,2 Early and late bacterial sepsis in the neonatal period often
involve different micro-organisms but both can be a major cause
of morbidity and, if untreated, carry a mortality of 50%.3 The
presentation of neonatal sepsis is often non-specific,4 although
acute onset with rapid, often catastrophic clinical deterioration
is common, especially in preterm infants and those with very low
birthweight (<1500g).5 Commonly isolated organisms include
coagulase negative staphylococci, Staphylococcus aureus,
Streptococcus agalactiae and Escherichia coli.
In contrast to the approach in older children, neonates are
presumed septic with minimal clinical suspicion.4 Consequently,
up to 75% of all neonates admitted to NPICU, Mater Dei
Hospital, Malta, undergo at least one septic screen, usually
including blood and cerebrospinal fluid cultures,6 immediately
followed by empiric broad spectrum antibiotics designed to
cover the more common organisms.7-9 Based on the prevailing
infecting organisms on our NPICU, benzylpenicillin and
cefuroxime are administered as empiric ‘first line’ antibiotics,
whereas co-amoxiclav and cefotaxime empirical ‘second line’
antibiotics if meningitis is suspected and meropenem and
teicoplanin as ‘third line’ if there is no apparent response to
the above antibiotics. Although a bacterial cause is confirmed
in less than 10% of these infants, antibiotics are continued for
at least 72 hours until cultures and indicators of sepsis (e.g.
C-reactive protein: CRP)6 are reported to be negative. Confirmed
sepsis necessitates a 7-14 day course of appropriate antibiotics,
according to the organism and site of infection.7
Until recently, it was Unit practice for neonates to be
observed in hospital for a further 24 hours once antibiotics
Malta Medical Journal Volume 21 Issue 04 December 2009
were discontinued, regardless of whether these were started
empirically or following positive cultures. This antibiotic-free
period was observed even though there were no indicators
of sepsis. However, although this policy was written into the
antibiotic-sepsis protocol for the Unit, it was not evidencebased and this study was designed to evaluate whether it is
clinically justified and, therefore, whether the practice could
be stopped safely.
Methods
All neonates aged >32 completed weeks gestation admitted
to the NPICU and who were started on antibiotics for suspected
or proven sepsis from December 2006 to January 2008 were
analysed prospectively. For all cases, data related to the
immediate antenatal history, the infant, investigations and
management were documented (Table 1). Adverse events were
defined as any events that occurred after stopping antibiotics
during the 24 hour ‘observation period’ and, if adverse events did
occur, all details relating to these events were documented. This
study set out to determine whether there was any relationship
between any patient data and discontinuing antibiotic therapy
and any observed adverse effects thereafter.
Premature infants aged <32 weeks gestation were excluded
from the study because, in this group, factors other than
antibiotic therapy (with or without an observation period after
stopping antibiotics), determine the time of discharge from
hospital (e.g. attainment of a gestational age of ≥35 weeks
and weight ≥2000g). Furthermore, once off all therapeutic
treatment, it is unit practice to encourage these babies to ‘room
in’ with their parents for between 1-3 nights prior to discharge.
Chi squared was applied to compare the association of any
adverse events during the observation period, taking p≤0.05
to be statistically significant.
Table 1: Data collected relating to patients,
investigations and management
General data
Maternal health during pregnancy, time of rupture of
membranes before delivery
Infant data
Date of birth, gender, gestation, birth weight, mode of
delivery, Apgar score at 1 and 5 minutes
age at admission to NPICU, temperature on admission,
symptoms and signs on admission
Investigations
White cell, neutrophil and platelet counts, sequential
CRPs, blood, urine and CSF cultures, chest X-ray
Management
Type of central line, mode and duration of respiratory
support (if any), any other support (e.g. inotropes)
type and duration (in days) of antibiotics
Malta Medical Journal Volume 21 Issue 04 December 2009
Results
During the study period, a cohort of 95 cases out of 297
admissions to the NPICU fulfilled the entry criteria. Of these, 69
were males and 26 females with a gestation of 32 to 43 completed
weeks (mean 34.75 weeks, median 38 weeks), and birth weight of
1.51 - 4.75kg (mean 2.97kg, median 2.95kg, SD 0.64). As shown
in Table 2, predisposing risk factors present in the mothers,
duration of ruptured membranes and mode of delivery were
not associated with any adverse effects on stopping antibiotics.
Similarly, no parameter relating to the infant’s gestation, birth
weight, Apgar score and clinical condition requiring admission,
investigations and management was observed to be associated
with any adverse events after stopping antibiotics.
On admission to NPICU, empirical first line benzylpenicillin
and cefuroxime were administered in 70 (74%) cases, coamoxiclav and cefotaxime (empirical second line antibiotic) in
21 (22%) cases, and 4 others received different antimicrobials
according to sensitivities. One infant included in this group also
received acyclovir for 14 days for suspected herpes infection. First
line antibiotic cover was subsequently changed to second line in
6 infants and changed to third line antibiotics (meropenem and
teicoplanin) in 1 infant, following a rise in CRP, growth in blood
cultures or deterioration in the clinical condition.
The duration of antibiotic treatment ranged from 2 to
14 days (mean 4.9 days, median 5 days, SD2.2). In 24 (25%)
and 53 (56%) cases, antibiotics were stopped after 72 and 120
hours, respectively, as serial CRPs and blood cultures remained
negative. Eighteen cases required antibiotics for longer periods,
ranging from 7-14 days, due to confirmed or strongly suspected
sepsis.
Of the 95 infants who received antibiotics, 26 (27%) infants
did not require any respiratory support, 13 (14%) needed oxygen
via nasal prongs, 52 (55%) had nasal continuous positive airway
pressure (CPAP) and 4 (4%) cases required intermittent positive
pressure ventilation.
Adverse events
After stopping antibiotics and prior to discharge from
hospital, neonates were observed without treatment on the
NPICU (52 cases) or postnatal ward (43 cases) for periods of
observation ranging from 0.5-13 days (mean 1.8 days, median
1 day, SD 2.2). The majority (73) were observed for 24 hours
before discharge. The length of observation was longer in 22
others due to other factors including prematurity, inadequate
weight gain and feeding difficulties.
No adverse events were recorded after stopping antibiotics
in any of the 95 neonates studied, including those who had had
a difficult neonatal course requiring ventilation and prolonged
courses of antibiotics because of documented sepsis (Table 2).
Discussion
Although several neonates are investigated for sepsis, only
3–8% will be found to have culture-proven infection, translating
into an incidence of just 2 in 1,000 live births.1,7 However, since
25
the early signs of sepsis in the newborn are often non-specific 4
and the consequences of untreated sepsis are often dire,3 many
newborns undergo diagnostic ‘septic screens’ and the initiation
of antibiotic treatment empirically, and certainly before the
diagnosis can be determined on culture.7-9 In our unit, this
practice is carried out in 75% of admissions to NPICU.
The risk factors for neonatal bacterial sepsis as listed in Table
3 are well documented.1,5,7 Many of these risk factors are common
and, therefore, empiric antibiotic use in neonates is justified.7-9
Equally important, however, is the discontinuation of these
antibiotics as soon as possible if sepsis is not confirmed and,
therefore, they are no longer required. Indeed, prolongation
Table 2: Association between cohort characteristics and risk factors with observed adverse effects
Documented parameter Details
Cohort (n=95)
Association with
adverse effects
• 69 male; 26 female
• gestation 32 to 43 completed weeks (mean 34.75, median 38 weeks)
• birth weight: 1.51 - 4.75kg (mean 2.97kg, median 2.95kg, SD 0.64)
None
Mothers (n=95)
• 5 group B streptococcal (GBS) colonisation; 4 fever during delivery;
2 foul smelling liquor (but culture negative); 1 genital herpes;
1 rubella seroconversion during pregnancy; 1 parvovirus infection
None
Rupture membranes
• 0 (elective LSCS) to 216 hours (mean 7.1hr, median 1.5hr, SD 24)
• four cases had prolonged rupture of membranes > 24 hours
None
Delivery
• 33 born vaginally, 5 ventouse, 25 elective LSCS, 32 emergency LSCS
• 82 had Apgar scores ≥7 at 1 & 5 minutes; 13 had score ≤6 at 1 or 5 min
None
Admission NPICU
• age at admission: 0.5 hours to 6 days (mean 8.5hr, median 2hr, SD 20)
• 62 babies normothermic, 30 hypothermic (<36.5oC), 3 febrile (>37.5oC)
• 68 (72%) respiratory distress; 9 (10%) with cyanosis,
6 vomiting and feeding problems, 5 omphalitis, 3 fever, 1 convulsions
None
Investigations
None
• total white cell range: 4.7-34.6x109/l with 19 cases abnormally raised
• neutrophils: 1.5-17.8x109/l with 10 cases having an abnormal count1
• 8 thrombocytopenia below <150 x 109/l
• CRP was positive (>10) in 11 cases (range 14-181)
Microbiology
• Positive blood cultures in 12 cases but only 2 with Serratia marcescens
and Staphylococcus aureus had raised CRP; other pathogens included
5 Staphylococcus albus*, 3 Enterococcus faecalis, 1 Moraxella osteolensis,
1 Group B Streptococcus (*NB: blood cultures collected by open drip method
accounting for number of presumed contaminants)
• CSF and urine cultures in 8 cases were all negative
• chest X ray showed increased hilar markings in 30 from total 90 cases,
confirmed to be within normal limits by reporting radiologists
None
Management
• 24 (25%) had indwelling central catheters (18 umbilical arterial,
5 umbilical venous, 2 radial arterial lines and 1 femoral venous line)
• benzylpenicillin + cefuroxime administered in 70 (74%) cases;
co-amoxyclav + cefotaxime in 21 (22%) cases; 4 others received
different antimicrobials according to sensitivities
• 1 infant included received acyclovir for 14 days for suspected herpes
• 6 infants received further antibiotics (meropenem + teicoplanin)
following a rise in CRP, growth in blood cultures or deterioration
in the clinical condition
• antibiotic duration 2 to 14 days (mean 4.9 days, median 5 days, SD2.2);
in 24 and 53 cases, antibiotics stopped after 72, 120 hours, respectively.
18 required antibiotics for 7-14 days, due to confirmed sepsis
• 26 infants needed no respiratory support, 13 nasal prong oxygen,
52 nasal CPAP, 4 cases required intermittent positive pressure ventilation None
Legend: SD=standard deviation; LSCS=lower segment caesarian section; CRP=C reactive protein; CPAP=continuous positive airway pressure
26
Malta Medical Journal Volume 21 Issue 04 December 2009
Table 3: Risk factors for bacterial sepsis in neonates
Newborn risks
Maternal risks
Prematurity
Premature rupture membranes Very low birth weight (<1000g)
Prolonged rupture membranes
Male>female
Maternal GBS
Low Apgar score (<6 at 1 & 5min)
Maternal fever >38.4oC
Perinatal asphyxia
Chorioamnionitis Meconium staining
Urinary infection
Congenital anomalies
Others
Poor antenatal care
Poor maternal nutrition
Substance abuse
Legend: GBS=Group B streptococcus
of unnecessary antibiotics has been shown to be associated
with, for example, greatly increased colonisation with gram
negative bacilli when antibiotics are continued for more than
3 days.10 Similarly, prophylactic antibiotic use to cover an
invasive device such as an intercostal drain, umbilical arterial
or venous catheter, does not prevent sepsis and is only likely
to select for multi-resistant organisms.11,12 Hence, if systemic
cultures are negative, antibiotics should be stopped after 48
hours. This is usually possible, particularly if automated blood
culturing systems are used, but delays of up to 5 days for a final
culture result are observed where this system is not established,
resulting in lengthier administration of empiric antibiotics.
Indeed, cessation of antibiotics at the earliest opportunity is
widely established and has not been shown to be associated
with a recurrence or the emergence of breakthrough sepsis.7
However, the additional ‘safety measure’ adopted by our Unit for
a further period of observation in hospital for potential adverse
effects is not evidence-based. Despite the study’s limitations in
a lack of a control group and failure to blind the authors to the
treatment the infants’ received, this study did not report any
adverse effects during the precautionary period of observation.
Indeed, this study analysed several known risk factors, itemised
above, that may have resulted in complications after stopping
antibiotics: yet, in all the cases, no events were recorded in the 24
hour period of observation, regardless of gender, weight, mode
of delivery, Apgar score, respiratory support, documented or
suspected sepsis, days and type of antibiotics received.
Conclusion
This study therefore challenges the current practice for a
period of pre-discharge observation, and strongly supports the
practice whereby neonates with a mean age of 38 weeks and
without other confounding problems, can be safely discharged
home on the same day as stopping antibiotics. As a result of this
study, the discontinuation of this unnecessary practice would
Malta Medical Journal Volume 21 Issue 04 December 2009
have reduced the average stay in hospital for 73 out of the total of
95 infants in the study cohort by 24 hours. This translates into a
cost saving on hospital stay of approximately €18,000 (including
bed stay and consumables) for this group of patients and, in turn,
a similar recurrent saving for the unit each year. Clearly, earlier
discharge would also minimise anxiety and inconvenience to
the family and, for all these reasons, the 24 hour precautionary
observation period has now been discontinued.
References
1. Rennie JM. Infections in the newborn. In: Roberton’s Textbook of
Neonatology, 2005, 4th Ed, London, Churchill Livingstone; 40.2: p.
521-45.
2. Hague KN. Definitions of bloodstream infection in the newborn.
Pediatr Crit Care Med. 2005;6(6):S45-9.
3. Kermorvant-Duchemin E, Laborie S, Rabilland M, Lapillonne A,
Claris O. Outcome and prognostic factors in neonates with septic
shock. Pediatr Crit Care Med. 2008;9(2):186-91.
4. Fischer JE. Physician’s ability to diagnose sepsis in newborns and
critically ill children. Pediatr Crit Care Med. 2005;6(6):S120-5.
5. Shah GS, Budhathoki S, Das BK, Mandal RN. Risk factors in early
neonatal sepsis. Kathmandu Univ Med J. 2006;4(2):187-91.
6. Isaacs D. Rationing antibiotic use in neonatal units. Arch Dis Child
Fetal Neonatal Ed. 2000;82:F1-F2.
7. Hutzal CE, Boyle EM, Kenyon SC, Nash JV, Winsor S, Taylor DJ,
et al. Use of antibiotics for the treatment of preterm partuition
and prevention of neonatal morbidity: a metanalysis. Am J Obstet
Gynecol. 2008;199(6):620e1-8.
8. Venkatesh MP, Garcia-Prats JA. Management of neonatal
sepsis by gram negative pathogens. Expert Rev Anti Infect Ther.
2008;6(6):929-38.
9. Arnon S, Litmanovitz I. Diagnostic tests in neonatal sepsis. Curr Opin
Infect Dis. 2008;21(3):223-7.
10.Isaacs D, Catherson J, Hope PL, Moxon ER, Wilkinson AR. Factors
influencing colonisation with gentamicin resistant gram negative
organisms in the neonatal unit. Arch Dis Child. 1988;63(5):533-5.
11. Lodha A, Furlan AD, Whyte H, Moore AM. Prophylactic antibiotics
in the prevention of catheter-associated bloodstream bacterial
infection in preterm neonates: a systematic review. J Perinatol.
2008;28(8):526-33.
12.Jardine LA, Inglis GD, Davies MW. Prophylactic systemic antibiotics
to reduce morbidity and mortality in neonates with central venous
catheters. Cochrane Database Syst Rev. 2008;23(1):CD006179.
27